Discussion of TreeletsAn adaptive multiscale basis for sparse unordered data.pdfVIP

The Annals of Applied Statistics 2008, Vol. 2, No. 2, 489–493 DOI: 10.1214/08-AOAS137F Main article DOI: 10.1214/07-AOAS137 c Institute of Mathematical Statistics, 2008 DISCUSSION OF: TREELETS—AN ADAPTIVE MULTI-SCALE BASIS FOR SPARSE UNORDERED DATA 8 By Catherine Tuglus and Mark J. van der Laan 0 0 University of California, Berkeley 2 l We would like to congratulate Lee, Nadler and Wasserman on u their contribution to clustering and data reduction methods for high J p and low n situations. A composite of clustering and traditional 5 principal components analysis, treelets is an innovative method for 2 multi-resolution analysis of unordered data. It is an improvement over traditional PCA and an important contribution to clustering ] P methodology. Their paper presents theory and supporting applica- A tions addressing the two main goals of the treelet method: (1) Un- . cover the underlying structure of the data and (2) Data reduction t prior to statistical learning methods. We will organize our discussion a t into two main parts to address their methodology in terms of each s of these two goals. We will present and discuss treelets in terms of a [ clustering algorithm and an improvement over traditional PCA. We 1